Epithelial ovarian cancer (EOC) is the most lethal of gynecologic malignancies. Advanced disease typically involves the upper abdomen and affects 70% of patients, associated with 5 year survival in the range of 10- 25% after treatment with surgery followed by chemotherapy. Early stage disease confined to the pelvis is associated with 5 year survival of greater than 90%, although the presence of high risk features still requires treatment with post-operative chemotherapy. Traditional clinical and molecular markers as stage, postoperative debulking status, p53 mutation, and BAX expression are reasonable but imperfect measures of outcome. This observation suggests that no single marker can serve as a surrogate for the complex genetic changes that are responsible for tumor growth and response to chemotherapy. In this regard, microarray gene profiling is a powerful technique that is capable of simultaneously assessing the expression of thousands of genes, although its clinical utility for patients with EOC remains to be determined. Using this technique, we have developed novel bioinformatics approaches to identify gene profiles in a training set that are highly prognostic of clinical outcome in EOC. In this grant, we will validate these data in a test set comprised of a large number of tumor samples from a separate institution, and we will also determine whether it is possible to streamline tumor profiling using RT-PCR and immunohistochemistry assays (Specific Aim 1). Furthermore, we challenge the generally accepted concept that accurate prognostic information can always be obtained from analysis of a static, pre-treatment tumor sample. Thus, in Specific Aim 2 we will obtain a dynamic assessment of gene expression in response to chemotherapy in vivo, based upon the accessibility of tumor cells from ascites immediately before as well as for several days after chemotherapy has begun. Finally, in Specific Aim 3 we will apply the micro-array technique to a study of patients with early stage disease, in an attempt to determine whether it is possible to identify only those patients who will derive the greatest benefit from post-operative chemotherapy. We anticipate that the ability to accurately identify predictive and prognostic factors in EOC will permit a more tailored approach to post-operative management for patients with this disease.